Hello Sniffer Folks, Version 2-3.0i5 has been posted for testing.
<http://www.sortmonster.com/MessageSniffer/Betas/MessageSniffer2-3.0i5-Distribution.zip> This version implements a number of performance enhancements relating to the use of a Persistent server instance. Be careful - this version has not been heavily tested. However it appears to be quite stable and extremely fast. MDaemon users should see a significant improvement in performance. Please let me know if this is true. The reason is that since most MDaemon system integrate Message Sniffer through the content filter and the content filter in MDaemon is apparently single threaded. The performance tuning features of this version allow the client and persistent server instances to coordinate much more closely with regard to the available computing power in the system so that polling delays _should be_ reduced significantly. _IN THEORY_ the improved signaling between client and persistent server instances will allow "polling synchronization" such that the highest possible performance for the hardware and load conditions can be achieved. Having said that, I hope it turns out to be true ;-) Let me know!! Multi-Threaded systems should see a significant improvement in performance also. In particular, jobs are now collected in batches by the persistent server instance. This eliminates a great deal of overhead (directory scans, locking, etc) on the part of the server instance. In addition, client instances can now detect the presence of a persistent server without locking and scanning the workspace. The combination virtually eliminates locking conflicts and any overhead usually associated with coordinating the peer-server process. I don't know precisely what performance increases you should expect from this. However I will say that in static tests the program ran so quickly that I thought it was broken. Message processing rates on my Toshiba 6100 were consistently above 14.2 Msg/Sec with CPU loads in the high 80s. I have not yet tested this version on *nix systems... but I expect similar performance improvements. I urge caution again - this is a somewhat radical set of changes with some unexpectedly good _apparent_ performance improvements. I'm not sure that I trust it yet... but I also haven't seen it fail. More testing needs to be done... (I appreciate your help!) Please do keep me informed. Thanks, _M Pete McNeil (Madscientist) President, MicroNeil Research Corporation Chief SortMonster (www.sortmonster.com) This E-Mail came from the Message Sniffer mailing list. For information and (un)subscription instructions go to http://www.sortmonster.com/MessageSniffer/Help/Help.html
